Classification of Engel knots (Q1751029)

An Engel 4-manifold is a smooth 4-manifold \(M\) equipped with an Engel structure \(\mathcal{D}\). The latter is a maximally non-integrable tangent 2-plane field, i.e., a \(2\)-distribution \(\mathcal{D}\subset TM\) which is locally spanned by vector fields \(X, Y\) such that \(X, Y, [X, Y]\) and \([X, [X, Y ]]\) frame the tangent bundle in the neighbourhood where \(X\) and \(Y\) are defined. The Engel line field, denoted by \(\ker(\mathcal{D})\), is the unique line field \(L\) defined by the conditions \(L\subset\mathcal{D}\) and \([L, [\mathcal{D},\mathcal{D}]]\subset[\mathcal{D},\mathcal{D}]\). A smooth embedding \(\gamma:S^1\to M\) tangent to \(\mathcal{D}\) is called an Engel (or a horizontal) knot in an Engel \(4\)-manifold \((M, \mathcal{D})\). Moreover, a formal Engel knot is a pair \((\gamma, (F_s)_{s\in [0,1]})\) consisting of a smooth embedding \(\gamma:S^1\to M\) and a homotopy \((F_s)_{s\in [0,1]}\) of injective bundle maps from \(TS^1\) to \(\gamma^\ast(TM)\) such that \(F_0=d\gamma\) and \(F_1(TS^1)\subset\gamma^\ast(\mathcal{D})\). Let \(\mathcal{E}(M,\mathcal{D})\) (resp. \(\mathcal{E}^f(M,\mathcal{D})\)) denote the space of all Engel knots (resp. all formal Engel knots) in \((M, \mathcal{D})\). The so-called scanning map is a forgetful map \(s:\mathcal{E}(M,\mathcal{D})\to \mathcal{E}^f(M,\mathcal{D})\) given by \(s(\gamma)=(\gamma,d\gamma)\). It is impossible for \(s\) to be a weak homotopy equivalence because closed orbits of the Engel line field \(\ker(\mathcal{D})\) are Engel knots. In the paper under review, the authors prove that the restriction of the scanning map \(s\) to the subspace of Engel knots which are not everywhere tangent to \(\ker(\mathcal{D})\) is a weak homotopy equivalence. This result can detect all data persistent under Engel homotopies, and hence essentially classifies all Engel knots.